Lunar and Planetary Science XLVIII (2017)
2857.pdf
Preservation of Tropical Subsurface Ice into the Very Recent History of Mars. M. R. Kirchof and R. E. Grimm.
Southwest Research Institute, 1050 Walnut St., Suite 300, Boulder, CO 80302. Email: [email protected].
Introduction: The history of volatiles on Mars and
implications for climate and geology are intimately
linked to the evolution of subsurface tropical ice. Ice is
currently not stable below the mid-latitudes [e.g., 1],
but the actual, time-integrated loss is uncertain [2, 3].
Layered-ejecta craters have long been thought to tap
buried ice [e.g., 4]. They are present at all latitudes and
sample to greater depths (kms) than possible with neutron spectroscopy or even surface-penetrating radar.
With the advent of near-global 10-m/pixel imaging of
Mars, formation model ages of individual layered
ejecta craters can be estimated from smaller craters superposed on their ejecta blankets [e.g., 5]. We focus on
single layered ejecta (SLE) craters because of their
prevalence at tropical latitudes. 50 craters from the
Robbins [6] database between ±30ºN and with diameters (D) ≥5 km have been analyzed so far. These
craters are chosen mostly from low dust regions (as indicated by Thermal Emission Spectrometer [7]) to
minimize effects due to dust deposition. Ages of these
craters provide new constraints on when and where
subsurface ice existed at tropical latitudes.
Methods: To estimate the formation model ages
of SLE craters, we measure small, superposed craters
(SSCs) on their ejecta blankets. The Neukum et al. [8]
production function is fit to the resulting SSC size-frequency distributions (SFDs) and their chronology
function is used to compute model ages. However, several issues introduce uncertainty in the age calculations: removal of SSCs by erosion and/or dust deposition, inclusion of craters only partially buried by the
ejecta blanket, inclusion of secondaries, and errors in
the chronology. While there is little we can do about
the last issue, we have developed some strategies to
mitigate the first three. Here we summarize these
strategies, which are discussed in detail in [9].
The first strategy is measuring craters of similar
sizes to the SSCs within a nearby reference area that is
on the same geological unit. Comparison of crater
SFDs for the two areas can indicate if any of these issues need to be considered. Our second strategy is to
compare these two area's SFDs to a subset of the SSC
SFDs only including degraded craters and obvious secondaries (those that form in chains and clusters). Similarities in these crater SFDs suggest that the SSC SFD
has been likely modified. Finally, the third strategy is
to evaluate production function fits to the SSC SFDs.
Diameter ranges that do not match within error are not
considered reliable for estimating model ages.
In the process of applying these strategies, we determine those model ages which appear to be the least
affected by crater removal, partially buried craters, and
secondaries. These are classified as “high confidence”
model ages, and are also analyzed for verification of
the full data set.
Once model ages are determined, we assess
whether the formation rate of equatorial SLE craters
has deviated from the formation rate of all low-latitude
craters. Our approach is to first determine the number
of SLE craters with model ages that fall into a given
0.5 Gyr bin (i.e., 0.25-0.75 Ga, 0.75-1.25 Ga, etc.).
These values are then normalized to the number of all
craters expected to form for the Neukum chronology in
each bin. A resulting plot of normalized number of
craters vs. age (e.g., Fig. 1) is used to assess if and how
the tropical SLE formation rate has deviated from the
background flux.
We also use the SLE crater model ages, along with
the craters' diameters to determine if and how depth to
the subsurface ice has varied with time.
Results and Discussion: Of the 50 equatorial SLE
craters examined so far, we ascertain that 29 of them
have high confidence model ages. Fig. 1 shows the
normalized number of craters vs. age for the high confidence set (results are similar for the full data set).
Data are currently insufficient to robustly conclude
how the SLE crater formation rate is changing with respect to the formation rate of all low-latitude craters.
The SLE cratering rate first appears to decrease from
~3 to 1 Ga and then increase again for the last billion
years, but uncertainties are still large.
If the trend were flat, as represented by the red line,
then the SLE crater formation rate has stayed the same
as the assumed rate for all low-latitude craters and the
amount of tropical ice has not considerably changed. If
the normalized number of craters decrease with decreasing age, then the SLE crater formation rate has
declined implying reduced tropical subsurface ice.
There is not, however, a straightforward explanation
why the SLE crater formation rate would increase towards the present. We suggest four possible explanations: 1) some SLE crater model ages are not formation
ages because of undetected removal of SSCs; 2) the
impact rate is higher for the last billion years than currently estimated by the Neukum chronology (as suggested by lunar studies [e.g., 10–12]); 3) SLE crater
ejecta is destroyed at a faster rate decreasing the number of older SLE craters that can be identified; 4) ice
has been recently cold-trapped at shallower levels due
to vapor diffusion resulting in the increased formation
of young SLE craters. Explanations (1-3) would mean
that the increase is not real, but a result of uncertainties
in the analysis, and would change if the biases were
quantitatively known and could be accounted for. Ex-
Lunar and Planetary Science XLVIII (2017)
2857.pdf
Table 1. High confidence model ages for equatorial
SLE craters likely formed in the Amazonian
Latitude (ºN) Longitude (ºE)
Figure 1. Number of tropical SLE craters with ages in the
given 0.5 Gyr bins (data plotted at center of the bin) normalized to the total number craters expected to form between
±30ºN for the Neukum chronology [8]. Error bars are Poisson.
Red line represents no change in formation rate of SLE
craters with respect to all low-latitude craters.
planation (4) would mean the increase is real and some
information is missing on the understanding of movement of subsurface volatiles. Overall, current limitations on the data and our understanding of crater erosion, movement of subsurface ice, and the chronology
inhibit fully resolving this issue.
Nevertheless, a key result from our estimation of
formation model ages is that SLE craters appear to
have formed throughout the Amazonian (Table 1).
Moreover, we have potentially found at least 2 SLE
craters that have formed within the last 500 Myr (up to
5 if lower confidence model ages are included). This
implies that tropical subsurface ice has been retained
even into very recent times.
Furthermore, the smallest SLE crater examined (of
the high confidence set) appears to likely be less than a
billion years old (D=5.8 km, -24.3ºN, 95.8ºE, age=0.7
–0.5/+0.8 Ga). Since excavation depth is related to
crater diameter (through the transient diameter; [13]),
this indicates recent tropical ice as shallow as 300400m in this location (Tyrrhena Terra). In addition, a
much older age for the second smallest crater (D=6.2
km, -27.8ºN, 3.41ºE, age=3.5 -2.5/+0.3) indicates ice
was this shallow even in the past, although it is in a
different area (Noachis Terra). Therefore, we can
broadly conclude that relatively shallow subsurface ice
has been present throughout most of Mars history, but
that the location may have varied through time.
D (km)
Age (Ga)
-7.803
86.01
8.85
0.3 -0.3/+0.8
-8.54
58.22
7.669
0.5 -0.5/+1.1
-28.35
271.95
8.943
0.6 -0.2/+1.0
-24.26
95.83
5.825
0.7 -0.5/+0.8
-1.6
350.1
10.011
0.8 -0.7/+1.4
-5.97
10.94
7.498
1.1 -0.6/+0.9
20.2
326.63
8.097
1.2 -0.8/+0.2
-27.39
148.51
7.863
1.5 -0.5/+0.6
-14.42
103.15
8.761
1.6 -1.0/+1.7
17.09
128.29
9.412
2.5 -1.4/+1.0
-25.9
152.41
8.958
2.5 -1.8/+1.1
-7.1
32.91
6.544
2.6 -2.6/+1.1
-18.13
2.22
10.182
2.9 -1.5/+0.7
Conclusion: Formation model ages have been
computed for 50 equatorial SLE craters using the density of smaller craters superposed on their ejecta blankets. For 29 of the estimated model ages, we have a
high confidence that the superposed crater SFDs are
likely not considerably altered by crater removal, partially buried craters, and secondary craters. Analysis of
these ages indicates SLE craters have formed at low
latitudes throughout the Amazonian (0-3 Ga), with a
few forming within the last 500 Myr. These results imply tropical subsurface ice has not been substantially
diffused away and is still present today. Moreover, this
ice has been regionally quite shallow, at least within
300-400 m of the surface, throughout Mars' history.
This work was supported by NASA grant
NNX14AM28G.
References: [1] M. T. Mellon et al. JGR 102, 19357–
19369, 1997. [2] S. M. Clifford & D. Hillel, JGR 88, 2456–
2474, 1983. [3] Grimm R.E. et al. JGR 121,
doi:10.1002/2016JE005132, 2016. [4] M. H. Carr et al. JGR
82, 4055–4065, 1977. [5] D. Reiss et al. MAPS 41, 1437–
1452, 2006. [6] S. J. Robbins & B. M. Hynek, JGR 117,
E05004, doi: 10.1029/2011JE003966, May 2012. [7] S. W.
Ruff & P. R. Christensen. JGR 107, 5119, doi:10.1029/
2001JE001580, Dec. 2002. [8] G. Neukum et al. SSR 96, 55–
86, 2001. [9] M. R. Kirchoff & R. E. Grimm. 47th LPSC,
2016, Abst. #1587. [10]
W. F. Bottke et al. Nature 449,
48–53, 2007. [11] N. E. B. Zellner et al. GCA 73, 4590–
4597, 2009. [12] M. R. Kirchoff et al. Icarus 225, 325–341,
2013. [13] H. J. Melosh, Impact Cratering: A Geologic
Process. 1989.